I have a situation where by I would like to 'pause' a health state alert (If thats the right phrase).
For example, if a Windows service stops, the health monitor detects this and raises an alert. This then generates a ticket in our ticketing system. The service restarts itself and the health state returns to healthy. The admin responding to the ticket finds nothing wrong.
Is there a way to get around this. I have been thinking about creating a custom monitor, but not sure how or what type. I am looking at something like Consecutive Samples over Threshold (But this is for perfomance).
Any ideas are welcome.
Thanks. Mark
Hi,
you can configure the monitor to not close itself when the health state returns to healthy (have a look at the alerting tab).
But even when the alert is automatically closed, it's still in the system. You could create a view to show all closed alerts from the last 24 hours or something like this.
cheers Stefan http://www.code4ward.net
Stephan's suggestions are both good ones. You also mentioned Consecutive Samples Over Threshold though. If you are looking to throw an alert (and thereby generate a ticket) only when the service has failed a certain number of times over a given time threshold, you can accomplish this through an event monitor. This is an old fashioned way of doing it, but it gives you greater control over what generates an alert which is very handy when you have a ticketing system connector.
Here's how you can do it: Create a repeated event monitor, look at the System log, Service Control Manager for the source, event 7036, filter description to include the name of the service to be monitored and "stopped". Then you can set the repeat thresholds (number of times repeated over what time threshold) to whatever you want to achieve the desired result.
-Rob